|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
86.137.23.129
I recently took time to evaluate the Cree Silicon Carbide diodes in a CD player [Marantz CD63KI] and I’ve come full circle with my opinion.The Silicon Carbide diodes replaced the IXYS HexFRED types which are a firm favourite, but Cree’s specification on the Silicon Carbide [10A/600V] tempted me to try them out in place of the IXYS.
My initial impression was very positive, a cleaner sound with improved definition across all frequencies.
After continued listening I started to find the sound somewhat fatiguing, too forward with a degree of steeliness or harness to the sound.
Before anybody jumps back and say’s the Cree diodes are revealing weaknesses elsewhere in the system, this is no OEM CD63, it sports a low jitter clock, Jung type regulators and a selection of premium capacitors in the analogue and digital sections.
So are the Cree diodes all that the spec sheets deemed them to be, is it a case of technology and sound quality not coming together?
At this point I’m only just holding off putting back the HexFRED’s so opinions please.
Follow Ups:
I would like to point out that Motorola schottky diodes (like MUR 860), 1N5822, 11DQ10 and 31DQ10 schottky diodes are not made of silicon carbide and sound VERY smooth to my ears, in several CD player and tube heater circuit mods.
you will also find several positive posts if you do a search in the Tweak asylum, Tube asylum or Bottlehead forum.What's being discussed is the sound of a quite recent development, high-voltage schottky diodes made by Cree.
You are possibly right in evaluating the positive and negative features of the Crees.
Unfortunatelly, I found further nougthy attributes of the Crees. A classical, 50A silicon diode bridge + 0.047 uF Jensen copper cap were replaced with a set of 25A Cree Silicon Carbid Schottkies in my tube power amp. This amp has the ability to handle tempo, rhythm in a very pleasurable way. Lifelike, dynamic sound became rather annoying, lifeless, artificial, calculable sound. Timing and energy peaks also belong to the week end of the Crees.But in a CD player, if the voltages aren't to high (and this is the usual case), why don't apply Motorola Schottkies, like MBR3045PT. Old, Made in USA or in Mexico, not newer ones produced in the Far East. Maybe this is the solution of your problem. If 30 Volt is not enough, you will find classical Schottkies up to 200 V, though these are newer production.
I'm courious of what annoying features will you find in the Hexfreds, when going back to the original configuration.
Hi KolbayThank you for your input here.
Your experience with the Cree diodes is similar to mine which is reassuring.
I have not tried lower voltage Motorola Schottky diodes but have used several types from International Rectifier. When exchanging the stock diodes for a 1.1A version I got the benefits of a cleaner sound, improved definition of notes but lost the musicality of the Silicon types I replaced.
The IXYS HexFRED are the only diodes I have used which give consistently excellent results.
I will not give up on the SIC diodes yet, perhaps I need to re-work the reservoir capacitor brand/type to swing back the balance??
The attribute of the SIC diodes I don't like is this 'steeliness' to the sound, the sound does not flow as well, it is too precise, over energetic and calculated.
I am starting to convince myself here that I am indeed on the wrong path with the SIC’s :0)
lovekylie
Not me.
But after bad experience with both sound and technical issues, I consulted an engineer friend of mine. The bottom line: see above.
As long as the devices are robust and used within their dissipation ratings, Schottky rectifiers should sound better than anything short of vacuum tubes.The problem with all solid-state junction rectifiers is minority carrier storage in the depletion region during forward conduction. When the input waveform reverses polarity, this charge is swept out in the form of a current pulse. These pulses have steep trailing edges, with lots of HF content.
The HF content in the reverse pulses sets the power supply transformer and associated wiring to resonating, with audible consequences to both the involved equipment and any other equipment sharing the power circuit or just located nearby.
Switch-mode power supplies lose significant efficiency if these pulses are large, as they are with conventional rectifiers. This is why the HexFRED and similar rectifiers were developed: they minimize the stored charge to improve the efficiency of switchers. Some audio people recognized HexFREDs' ability to reduce the magnitude of the input noise to the power supply issue and adopted them more than a decade ago. "Minimize" does not mean eliminate, however.
Schottky rectifiers are majority carrier devices. They do not store charge, so there is essentially no reverse pulse when the input waveform reverses polarity. Silicon Schottky rectifiers are limited to low peak reverse voltages, so they traditionally have been used only for filament DC supplies and similar low-voltage applications. Their lower forward voltage drop compared to silicon junction rectifiers is the main reason for their use.
CREE developed silicon carbide Schottky rectifiers a few years ago. Silicon carbide is a semiconductor with a larger band-gap than silicon, so the Schottky diodes can be made with higher peak reverse voltage ratings. The trouble with silicon carbide is that it is a refractory material, and difficult to grow into defect-free single crystals. If there is a crystal defect within the rectifier, it may fail under normal service. I recall hearing about some failures when these devices first became available, but I don't know if CREE has found test procedures that would guarantee reliable operation of parts sold today.
Hello Al.I also wish I could through a technical argument in, The lecture I got from my engineer pcb's designer friend was so long and educated, beyond my ability to just forward it from memory, I hope I can make it next week.
Anyway, being a humble sound technician for more than 20 years, I trust my ears more than any technical explanation or scope measurements, we all know some electronic parts measure the same and sound different.
I can't agree more with lovekyle pin-point description of Schottkys sound, Every bridge rec's I've tried to put them in, by-passed or not, I got the same result he was talking about.
My discussion was theoretical only, as I've not experimented with the silicon carbide Schottky rectifiers.Two thoughts as to why they might give the results you report:
First, perhaps a little rectifier noise helps hide other problems in the circuits or external noise environment, so that the SiC Schottkys are actually revealing things that were always there.
Second, perhaps the SiC Schottkys have some forward impedance differences compared to silicon junction devices. They should be less temperature sensitive, for one thing.
Hi Al.You say: "My discussion was theoretical only, as I've not experimented with the silicon carbide Schottky rectifiers".
Well, That's the whole point, is seems that the Schottky hype cloud is settling down and negative comments are starting to show up. But that goes mainly for subjective audible impressions.
Technically (again, educated arguments to arrive in the near future) Schottkys are not suitable for audio bridge rec's, I've used 11DQ10, with specs more than adequate for low voltage cdp's/dac's and each time the regulators boiled compared to before, and I have tried better bigger caps and transformer.
You referred to component weaknesses emerging because of lower noise floor and other technical merits that Schottkys posses, sorry, can't agree here as well from a subjective point of view. Most of the tweaks that do this are, IMHO, the addition of a clock, p.s. caps bypassing, dedicated rails and all-over better AC mains route.
"Technically (again, educated arguments to arrive in the near future) Schottkys are not suitable for audio bridge rec's, I've used 11DQ10, with specs more than adequate for low voltage cdp's/dac's and each time the regulators boiled compared to before, and I have tried better bigger caps and transformer."Either the rectifiers are working or they are not. The small additional DC voltage that results from using Schottkys instead of junction rectifiers should not cause regulators to "boil."
I hope you can clarify this, because others I know have used low voltage Schottkys successfully.
Maybe this will shed some light.
- http://db.audioasylum.com/cgi/m.mpl?forum=tweaks&n=122636&highlight=carlos+schottky&r=&session= (Open in New Window)
If the regulators are used at the limits of their capabilities, then a few hundred millivolts increase in input voltage may fry them.This consideration should not be a problem for output stages, where the supply rectifiers handle larger currents and generate more noise, and the supply is typically not regulated.
Hello Al SekelaI appreciate your technical imput here and everything you say about the Silicon diodes ability to eliminate (well 99%) switching noise is true.
I only wish I could through a technical argument in as to why so the HexFRED diodes are preffered and perform better overall to the SIC.
Sure the SIC diodes outperform the FRED's in terms of dynamics, leading edges, lowest noise floor; but with this I am detecting a degree of hardness, forwardness, and unatural sheen to the mid-high frequencies.
Further testing on my part needs to be made so I’ll switch back to a new sets of FRED’s for the analogue stage only and evaluate the sound.
I’ll report my findings back here.
Did you bypass the Crees ? You should install a 0.01 uF stacked film accross each diode i.e. from the cathode to the anode. I haven't tried them as yet because my system is under revamping. However, several audiophile friends, whose ears I trust say they are better than the GaAs diodes, which were king of the hill untill the Crees arrived. One audiophile friend who is also an E.E. (oxymoron ?) said "If you want to wring the last iota of performance from a unit use the Crees.
Thanks pkell44No I haven't tried this, what type of capacitor do you suggest?
So this brings me to the questions, do the Cree SIC diodes come with a degree of forwardness, hardness which should be counteracted with capacitors?
Since I never had this issue with the HexFRED I never felt compelled to add any capacitors they just worked from day one.
This post is made possible by the generous support of people like you and our sponsors: